Some of the best ways to lower Azure spend are common-sense approaches to cloud management, whereas others require a little more thought. Our list of 10 Ways to Lower Azure Spend combines a mixture of both, and we also suggest one way in which you can keep Azure spend permanently lower.
Businesses migrating to the cloud often find their cloud spend is much higher than anticipated. This isn’t necessarily through a lack of planning or even operational governance. Often it’s due to businesses being unaware of the many different price components that contribute towards a cloud bill and the fact that—in the cloud—you pay for what you provision rather than what you use.
For example, you may believe you’re provisioning a Virtual Machine (VM) with the right amount of compute, memory, and storage capacity. However, there are likely other components attached to the VM that are increasing your cloud spend. Then, when you terminate the VM, if you’re unaware of the other components, they continue to accrue costs without your knowledge—until you receive your bill!
Microsoft Azure is one of the worst offenders for adding components to a VM—and it doesn’t make it easy to find them when you get your bill. So, we have compiled a list of 10 ways to lower Azure spend and produced a free eBook you’re invited to download which expands on each of the items on our list and includes pro tips for best cost management practices.
10 ways to lower Azure spend
Before applying these 10 ways to lower Azure spend, it’s important to have total visibility across your cloud environment. Microsoft claims its Azure Monitor portal provides “all the visibility you need,” but although this is a decent tool for accessing utilization data on a shallow level, identifying some issues, and seeing opportunities to optimize performance, it doesn’t really go far enough for most businesses.
Ideally, it’s better to use a third-party solution such as CloudHealth that provides single-pane visibility of all your assets and the components attached to them. With CloudHealth, you don’t actually have to look very hard to lower Azure spend because the platform’s “Health Check” capability will flag cost-saving opportunities automatically—among which you’ll likely find:
Unattached disc storage
When you launch a VM, disc storage is usually attached to your application. When you terminate the VM, the disc storage remains active and you continue paying for it even though you’re not using it. Any disc storage that’s been unattached for more than (say) two weeks should be terminated.
Realistically, how many snapshots do you need to facilitate a recovery if a VM fails? Microsoft allows you to configure settings to automatically create snapshots but not to delete older snapshots. Aged snapshots are usually obsolete but cost money to keep. These need to go!
The term “zombie assets” is often used to cover every unused asset in cloud computing, but you need to pay careful attention for VMs that may have been abandoned after failing to launch properly and IP addresses that were previously attached to VMs but are now longer being used (but charged for).
Older generation VMs
If your business has only recently migrated to the cloud, it’s unlikely you’ll have any VMs running on Azure Classic, but more mature businesses will want to search out any older generation VMs and upgrade them to the more cost-efficient Azure Resource Manager (ARM) VMs.
VMs suitable for rightsizing
As mentioned earlier, when you operate in the cloud, you pay for what you provision. So, if you provisioned a VM with a little “wiggle room”, or the utilization of the VM has subsequently decreased, now is the time to downgrade it to a size befitting its capacity requirements.
Disc storage suitable for rightsizing
If you provision a VM without stipulating the type of disc storage you require, Microsoft will attach premium disc storage by default. If you don’t need premium disc storage for your application, get rid of it. You could be paying up to three times the price of standard storage.
SQL Databases suitable for right sizing
Similar to disc storage, Microsoft Azure offers different tiers of SQL Databases—Basic, Standard, and Premium for high-performance applications handling large volumes of requests simultaneously. If your application doesn’t get a large volume of requests, move it down to an appropriate tier.
VMs you can start and stop on a schedule
Non-production VMs used for developing, testing, staging, and QA don’t need to be left running 24/7/365. By scheduling “on” times for 8:00 a.m. to 8:00 p.m. Monday to Friday, you’ll save nearly 65% of the cost of running these assets—or more if you apply a more aggressive schedule.
VMs suitable for Reserved VM Instance purchases
If you have VMs on Pay-as-You-Go billing that are running more than 400 hours per month, they’re likely candidates for a Reserved VM Instance. You can lower Azure spend by up to 72% on these VMs depending on their size, their type, and the number of hours they’re running each month.
Infrequently accessed data
Microsoft offers a combination of data storage options depending on your desired level of redundancy and how often the data is accessed. If you find you’ve infrequently accessed data stored in a “Hot” storage volume, it’s good business sense to move it to a less expensive option.
One way to keep Azure spend lower
Our eBook concludes by reminding readers that the 10 ways to lower Azure spend aren’t one-time or periodic exercises but need to be ongoing processes. “Ongoing processes” could be interpreted as “as soon as you reach number 10, start again at number one,” but that’s not the intention. What the eBook is recommending is that you implement a solution that monitors your business’s activity in the Azure Cloud and alerts you of opportunities to lower Azure spend when they occur.
It was mentioned above how CloudHealth’s “Health Check” capability can flag cost-saving opportunities automatically, but it’s actually smarter than that. CloudHealth can be configured to monitor your Azure Cloud environment around the clock and notify you when it finds unattached disc storage, aged snapshots, or zombie assets suitable for terminating. Truthfully, you can even configure the platform to initiate a function to terminate unused assets itself.
This process is known as “policy-driven automation,” in which you create a policy and CloudHealth will tell you (or act) when the policy is violated. For example, if you configure CloudHealth to alert you to VMs suitable for rightsizing, it will send you an email whenever the utilization of a VM has fallen below your defined threshold. It’s effectively hands-free management, and it has many practical uses for lowering Azure spend, optimizing performance, and enhancing cloud security.
Find out more about keeping Azure spend lower with automation
Hopefully, our 10 ways to lower Azure spend eBook will be of value to you and will help you find opportunities to reduce your business’s cloud costs. To help further explain how policy-driven automation can help lower Azure spend permanently, you’re invited to see a demo of CloudHealth in action and take a free trial of our cloud management platform to evaluate its automation capabilities in your own environment. There’s no obligation required before you can take advantage of this opportunity, so contact us today to find out more.